The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2016-11-21

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:04 polettix joined #mojo
00:34 trwww joined #mojo
00:45 jabberwok joined #mojo
01:40 aborazmeh joined #mojo
01:43 Paddi joined #mojo
02:11 inokenty-w joined #mojo
02:35 ssm joined #mojo
02:45 jberger Maybe that can be the new escape key!
02:46 * jberger is pretty exhausted from Meta::Hack, that joke doesn't make much sense
03:34 noganex joined #mojo
03:54 Dandre joined #mojo
04:21 dave joined #mojo
04:47 dave joined #mojo
04:56 _dave_ joined #mojo
05:04 dave__ joined #mojo
05:04 dboehmer_ joined #mojo
05:43 dave joined #mojo
05:54 dave joined #mojo
06:05 dave joined #mojo
06:08 Paddi joined #mojo
06:19 aborazmeh joined #mojo
06:26 dod joined #mojo
06:30 dod joined #mojo
07:10 dave joined #mojo
07:16 Vandal15263 joined #mojo
07:27 polettix joined #mojo
07:32 dave joined #mojo
07:35 dod joined #mojo
07:38 dave joined #mojo
07:39 mbudde joined #mojo
07:51 AndrewIsh joined #mojo
07:51 dod joined #mojo
08:04 ledet joined #mojo
08:05 jamesaxl joined #mojo
08:17 salva joined #mojo
08:18 trone joined #mojo
08:19 janl joined #mojo
08:33 dod joined #mojo
08:38 khfeng_ joined #mojo
09:16 Paddi joined #mojo
09:38 dod left #mojo
09:43 rshadow joined #mojo
09:50 dod joined #mojo
09:59 osfabibisi joined #mojo
10:03 jontaylor joined #mojo
10:32 jontaylor joined #mojo
10:36 laidback_01 joined #mojo
11:06 tchaves joined #mojo
11:11 parv joined #mojo
11:12 tchaves1 joined #mojo
11:17 tchaves joined #mojo
11:31 coolo joined #mojo
11:40 sri anyone got an ikea signum? any good?
11:41 pink_mist oh one of those, yeah I have one
11:41 pink_mist it's not bad
11:41 sri thinking about doing some proper cable management :)
11:41 janl under table cable trays works as well as you make them work
11:41 pink_mist yeah, what janl said =)
11:41 janl get some cable ties too
11:42 sri already got velcro cable ties
11:42 janl ikea had velcro cable ties when I last looked 15 years ago.
11:43 sri it's hard to do cable management with two laptops on the desk
11:44 sri don't want to drill holes into the desk ;p
11:51 mishanti1 sri: I love using either velcro cable-ties with adhesive for under-desk mounting, or good old cable-ties in combination with cable tie anchors.
11:52 mishanti1 Has helped avoid the inevitable ball of wires that those trays tend to produce. Also what janl wrote: the trays work as well as you make them work, so if you are tidy then that might serve just as well.
11:53 sri btw. the dell wd15 docking station is absolutely horrible
11:53 janl I am tidy under duress
11:54 sri it's like they designed it without having access to the laptop that would later use it
11:55 sri they actually put two usb ports on the front of the thing, but placed the cable that would connect to the laptop in a way that only allows you to place the docking station way behind the laptop
12:03 gregf_ joined #mojo
12:42 jontaylor joined #mojo
13:08 alexbyk joined #mojo
13:26 dotan_convos joined #mojo
13:29 gizmomathboy joined #mojo
13:39 bwf joined #mojo
13:50 gryphon joined #mojo
14:06 jabberwok joined #mojo
14:10 * jberger hugs mojo's centralized router
14:11 jberger 4 days of working in catalyst has me baffled and anxious
14:12 jberger I probably would get it after a while, but 4 days wasn't enough
14:13 Pyritic joined #mojo
14:22 [vlad] jberger: in cat you can set $c->response->body basically all the response stuff is in c->response
14:36 marty joined #mojo
14:41 khfeng joined #mojo
15:00 gryphon joined #mojo
15:08 jberger [vlad]: that isn't the problem, the routing paradigm is completely different
15:08 jberger once I end up in a controller/action I'm fine
15:11 jberger (well, fine-ish)
15:11 * [vlad] is trying mojo
15:11 [vlad] cat didn't seem to want to support websockets nicely
15:12 jberger enjoy!
15:12 jberger no, in some sense websockets are why sri started Mojolicious
15:12 jberger (this is a very general statement, but I think it is somewhat true)
15:31 orev joined #mojo
15:49 sh14 joined #mojo
15:57 Paddi joined #mojo
16:00 HtbaaPi is there any way to debug why hypnotoad is no longer working as being told?
16:00 HtbaaPi I start with hypnotoad -f script/myapp
16:00 HtbaaPi but instead of running in the foreground I get a "Starting hot deployment for Hypnotoad server 1" and it exits
16:01 HtbaaPi had no issues earlier today (Vagrant + Docker environment). When I run the app through morbo there are no issues
16:01 janl joined #mojo
16:01 HtbaaPi and running with the --test flags I get a "Everything looks good!" message
16:02 HtbaaPi I do have some warnings about given/when being experimental, but those were there all day
16:02 janl might be there all year
16:03 kes joined #mojo
16:04 HtbaaPi janl: perl 5.24 does that yes
16:05 polettix joined #mojo
16:06 HtbaaPi anyway, the same behavior is being shown when not running in a docker container. I've really got no clue as to why hypnotoad isn't respecting the -f flag
16:06 jberger HtbaaPi: did you upgrade mojo recently? is your environment cleanish? are you passing arguments to the start command in your script?
16:07 HtbaaPi it's running the most recent Mojolicious yes. Environment is clean. Built a fresh container image for it. Arguments are being passed properly as well
16:08 HtbaaPi the other flags such as -h and -t do work
16:09 jberger you are running the most recent mojolicious, but have you upgraded since it worked?
16:09 HtbaaPi could be, I rebuilt the image today several times.
16:11 HtbaaPi downgrading Mojolicious doesn't make any difference it seems
16:11 jberger what did you downgrade it to?
16:11 jberger try 7.09 or less
16:11 HtbaaPi just did, no difference
16:12 HtbaaPi I even downgraded back to 6.31 (which it was running on before) and the same thing happened
16:12 jberger and just to confirm, you aren't munging @ARGV or passing anything but the name of the application to ->start_app, right?
16:12 HtbaaPi I tried rebooting the Vagrant VM a couple of times but again no difference
16:12 HtbaaPi hypnotoad -f script/myapp
16:12 HtbaaPi that's all
16:13 HtbaaPi the script is the default
16:13 CHYC HtbaaPi: May not be relevant to you, but when I start hypnotoad at boot using systemd and then run "hypnotoad -f script/myapp", I get your symptoms described. I don't use containers, so it's not PID 1.
16:13 CHYC So basically it's when I have the app running already.
16:13 jberger oh, if you have multiple instances of the application running they will need different pid files
16:14 HtbaaPi just a single instance
16:15 HtbaaPi perl script/myapp daemon works fine as well. But not hypnotoad
16:16 jberger can you dump your env and confirm that there are no HYPNOTOAD_ or MOJO_ variables in there
16:17 HtbaaPi no such variables
16:17 jberger ok
16:17 jberger I'm about out of ideas myself
16:18 jberger are you sure that the hypnotoad script you are using is the correct one and has the right #!
16:18 HtbaaPi yes
16:18 HtbaaPi whoah... This is strange!
16:18 HtbaaPi in my dev env I'm running as root, mainly because of docker
16:18 HtbaaPi so I su vagrant and tried running it locally and no issues
16:19 HtbaaPi wow this is really strange
16:19 jberger systemd?
16:19 HtbaaPi it's Ubuntu 16.04 so yes
16:19 HtbaaPi as user vagrant running docker-compose up it all works again
16:19 jberger process management on systemd is ... something I'm not going anywhere near
16:20 jberger maybe someone else has ideas
16:20 HtbaaPi oh wait...
16:21 HtbaaPi nope, still issues
16:22 HtbaaPi even when running through sudo
16:22 HtbaaPi luckily the work day has come to an end. But sure would've been nice if I could have finished my work today. Oh well, can't have it all :-)
16:22 HtbaaPi thanks for the help anyway. It's much apreciated
16:24 jberger good luck!
16:24 purl You'll need it.
16:24 HtbaaPi thanks
16:25 HtbaaPi I am running it through the prefork server now and I'm getting an error from an undefined error in a template
16:25 HtbaaPi reason that didn't show up with morbo is because that template is only being called in a part where the app is doing HTTP requests to itself. Which morbo doesn't support, hence that part was being tested with hypnotoad :)
16:25 HtbaaPi lets see if fixing this will actually fix this issue
16:26 jberger oooh, I really dislike apps calling back into themselves
16:26 jberger is it possible for you to extract that bit into a reusable helper?
16:26 lluad joined #mojo
16:27 jberger I've only really needed that once and it was a very specific thing (using wkhtmltopdf to get a pdf representation of a page)
16:27 jberger and it still caused all kinds of problems
16:28 HtbaaPi Well, that's exactly what I'm doing: using wkhtmltopdf
16:28 HtbaaPi but it actually does API calls to itself as well
16:29 * jberger shivers
16:29 HtbaaPi I'm not too happy about that part though
16:30 janl recursion is devine
16:32 HtbaaPi hmm, even though those errors have been fixed hypnotoad still refuses to start
16:33 HtbaaPi or at least when executing docker-compose up as root
16:36 HtbaaPi at least running it with sudo it works and that's good enough for me. In production hypnotoad isn't being run as root either
16:36 jberger that still sounds like an environment issue
16:37 jberger what is in your non-root environment?
16:38 jberger sudo uses root permissions but your user environment
16:38 HtbaaPi nothing out of the ordinary. It's just the default Ubuntu 16.04 image from Vagrant
16:38 HtbaaPi I do a sudo tmux and go from there
16:39 jberger tmux as root :o
16:39 jberger that sounds very strange
16:39 jberger anyway, we are passed my ability to help
16:39 jberger the vagaries of process management are still not my forte
16:40 HtbaaPi the only thing that was out of the ordinary was that I was tailing the output of one of the logfiles (docker-compose exec app tail -f log/production.log) and that terminal crashed but I couldn't kill it either
16:40 HtbaaPi after that hypnotoad would no longer run
16:40 jberger but I might not be very smart, I can't even rename a controller in Catalyst and have it keep working :(
16:40 HtbaaPi :)
16:40 HtbaaPi I haven't touched Catalyst in years
16:40 HtbaaPi wait that's not actually true
16:41 HtbaaPi I had to move stuff from an old Ubuntu 7.04 server to a new 16.04 and it had an very old Catayst app that I developed
16:41 HtbaaPi I uploaded it to the new server, even installed the latest Catalyst and it worked like a charm. Had me surprised :)
16:41 jberger HtbaaPi: is there an existing pid file that needs to be cleaned up?
16:41 HtbaaPi ps: I only run tmux a root when in a vagrant VM :)
16:42 jberger or perhaps a log file with the wrong permissions (that can't be opened)?
16:42 HtbaaPi oh good one!
16:42 purl You know it!
16:42 lluad joined #mojo
16:42 HtbaaPi that could probably it
16:43 HtbaaPi because now it is running as root after I deleted the pid file in the scripts directory
16:43 jberger sounds like it
16:43 HtbaaPi it being a dev environment I add most of the current directory as a volume to the docker container, and also the scripts directory
16:43 HtbaaPi wow
16:43 HtbaaPi at least I'm glad we figured that out!
16:44 jberger \o/
16:44 HtbaaPi should explain why after that crash I couldn't start Hypnotoad anymore
16:44 HtbaaPi nice insight :)
16:46 rshadow joined #mojo
16:52 jberger interestingly, it is actually as if you were trying to run two copies of the same application (as I mentioned before)
16:52 stryx` joined #mojo
16:52 jberger but you weren't doing so intentionally
16:52 jberger I'll have to remember that one
16:52 HtbaaPi couldn't it just have been because the pid file remained?
16:53 janl yes
16:54 jberger yes, but that's basically what would happen in the case of two apps (which don't use different pid file names)
16:54 jberger nevermind, not important
16:54 HtbaaPi :)
17:02 jberger :o
17:03 jberger http://news.perlfoundation.org/2016/11/the-perl-conference-save-date-2017.html
17:07 stryx` joined #mojo
17:08 HtbaaPi I'm calling it a day. Thanks for the help jberger and CHYC
17:08 HtbaaPi cya
17:09 jberger cheers
17:29 PryMar56 joined #mojo
17:36 dboehmer joined #mojo
18:20 FatalNIX jberger's always got it covered
18:24 * jberger blushes
18:29 pink_mist jberger: is it a known problem that Mojo::ACME doesn't seem to want to work with multiple domain names for a single cert?
18:29 pink_mist (i.e. www.xxx.yyy and xxx.yyy)
18:29 jberger it should
18:29 jberger but I don't know for sure if I've actually tried it
18:30 pink_mist $ ./r-d.pl acme cert generate -t -n test -a account-t.key www.rainbow-da.sh rainbow-da.sh
18:30 pink_mist [Mon Nov 21 18:40:58 2016] [debug] Reading configuration file "/home/suppy/sites/r-d/r-d.conf"
18:30 pink_mist Can't call method "begin" on an undefined value at /home/suppy/perl5/lib/perl5/Mojolicious/Plugin/ACME/Command/acme/cert/generate.pm line 56.
18:30 pink_mist No cert was generated
18:30 jberger hmmm, well, that certainly doesn't look right :-P
18:31 pink_mist if I only supply one domain name it works
18:31 pink_mist (either one)
18:34 jberger oh carp
18:34 jberger its an easy one
18:34 pink_mist excellent :P
18:34 jberger pink_mist: since you are all setup to test can you try this out for me?
18:34 jberger https://github.com/jberger/Mojo-ACME/blob/master/lib/Mojolicious/Plugin/ACME/Command/acme/cert/generate.pm#L56
18:35 jberger can you replace that with
18:35 jberger sub { my $delay = shift; $acme->new_authz($_ => $delay->begin) for @new },
18:35 * jberger facepalms
18:35 purl facepalms are certainly another important measure
18:36 pink_mist sure :P
18:36 * jberger pushes purl down the stairs
18:36 * purl attaches herself to jberger using suction and a ham straw
18:37 pink_mist yay! that works! =) (let me just check if the actual cert works right too)
18:38 jberger it becomes clear how much more work I put into the library than the script :s
18:39 pink_mist yep, works brilliantly =) thanks :P
18:39 jberger sweet
18:39 jberger I'll quick patch that and upload
18:46 dod joined #mojo
18:47 jberger released
18:47 jberger hmmm is cpan_mojo dead?
18:48 jberger also, I wonder if the recent metacpan work distrupted it
18:48 jberger disrupted even
18:49 cpan_mojo Mojo-ACME-0.08 by JBERGER https://metacpan.org/release/JBERGER/Mojo-ACME-0.08
18:49 jberger \o/
18:49 pink_mist there we go :)
18:49 jberger wow, that took a while
18:49 pink_mist jberger++
18:50 jberger I guess it did just show up on metacpan recent
18:55 jberger pink_mist: thanks for the report
18:56 jberger I really need to find a way to improve the test coverage on the script itself (ie any coverage)
18:56 preaction if y'all want, i could hack some kind of websocket notification broker for cpan uploads
18:57 preaction it's part of my long-term plan for cpantesters, so i'd just be skipping a few steps to do for just the uploads
18:57 jberger batman hosts cpan_mojo now
18:57 jberger batman: does that sound useful?
18:58 preaction i mean, either way i'mma do it. just hack that in to the existing data scripts to feed via Mercury as part of the API project. then i can move to making a new Metabase API (and a legacy one for the transition period)
19:00 preaction also, it'd be nice if i could openapi those websockets... i totally forgot i changed mercury to be embeddable in existing mojolicious apps
19:03 preaction ... i swear i'm going to spend the next week just finalizing and fleshing out all the to-do list items i've come across because of the hackathon...
20:10 stryx` joined #mojo
20:11 dod joined #mojo
20:44 jontaylor joined #mojo
21:13 marty_ joined #mojo
21:26 disputin joined #mojo
21:48 Adurah joined #mojo
21:56 rshadow joined #mojo
22:19 eseyman joined #mojo
22:20 polettix joined #mojo
22:50 Zx3 joined #mojo
23:39 stryx` joined #mojo

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary